Abstract
Background:
Nearly 900 000 people in the United States are living with diagnosed human immunodeficiency virus (HIV) infection and therefore increased cancer risk. The total number of cancers occurring among HIV-infected people and the excess number above expected background cases are unknown.
Methods:
We derived cancer incidence rates for the United States HIV-infected and general populations from Poisson models applied to linked HIV and cancer registry data and from Surveillance, Epidemiology, and End Results program data, respectively. We applied these rates to estimates of people living with diagnosed HIV at mid-year 2010 to estimate total and expected cancer counts, respectively. We subtracted expected from total cancers to estimate excess cancers.
Results:
An estimated 7760 (95% confidence interval [CI] = 7330 to 8320) cancers occurred in 2010 among HIV-infected people, of which 3920 cancers (95% CI = 3480 to 4470) or 50% (95% CI = 48 to 54%) were in excess of expected. The most common excess cancers were non-Hodgkin’s lymphoma (NHL; n = 1440 excess cancers, occurring in 88% excess), Kaposi’s sarcoma (KS, n = 910, 100% excess), anal cancer (n = 740, 97% excess), and lung cancer (n = 440, 52% excess). The proportion of excess cancers that were AIDS defining (ie, KS, NHL, cervical cancer) declined with age and time since AIDS diagnosis (both P < .001). For anal cancer, 83% of excess cases occurred among men who have sex with men, and 71% among those living five or more years since AIDS onset. Among injection drug users, 22% of excess cancers were lung cancer, and 16% were liver cancer.
Conclusions:
The excess cancer burden in the US HIV population is substantial, and patterns across groups highlight opportunities for cancer control initiatives targeted to HIV-infected people.
Nearly 900 000 people were living with diagnosed human immunodeficiency virus (HIV) infection in the United States at the end of 2010 (1). Among HIV-infected people, cancer risk is increased because of immunosuppression, frequent coinfection with oncogenic viruses, and risk behaviors including smoking (2–4). Since the introduction of highly active antiretroviral therapy (HAART) in 1996, mortality related to advanced immunodeficiency (ie, acquired immunodeficiency syndrome [AIDS]) has declined (5,6), leading to increasing numbers of people living with HIV and an overall aging of the HIV population (7).
Three virus-related cancers are considered by the Centers for Disease Control and Prevention (CDC) to define the onset of AIDS in an HIV-infected person: Kaposi’s sarcoma (KS), non-Hodgkin’s lymphoma (NHL), and cervical cancer. The incidence of these AIDS-defining cancers (ADCs) has decreased since the introduction of HAART, but the incidence of some non–AIDS defining cancers (NADCs) has increased (8). Due largely to the increasing size and advancing age of the HIV population, the number of NADCs has risen, leading to growth in the overall cancer burden among HIV-infected people (7).
The burden of cancer in a population can be described in terms of the number of cases, both overall and attributable to specific exposures such as infections, alcohol, and excess body mass index (9–12). Such estimates allow cancer and its risk factors to be understood in a broad context to inform research and public health planning. Until recently, incomplete national HIV surveillance data prevented estimation of the cancer burden in the HIV-infected population for the entire United States (7). Some cancer cases are expected in the HIV population, because of background risk alone, and these can be estimated using cancer incidence rates from the general population. However, given higher risks, larger numbers of cases would be expected in the HIV population for many cancer types. These excess cancers highlight differences in the cancer burden between HIV-infected people and the general population, providing one framework for considering how cancer control strategies can be tailored to HIV-infected people. In principle, the excess cancer burden represents cases that could be eliminated if one could intervene on all aspects of being HIV-infected, thus reducing cancer risks to those in the general population.
Defining optimal strategies for cancer prevention and early detection in this growing, high-risk population with unique public health needs is a rising challenge (13). To provide context for research and public health efforts, we quantified the total and excess cancer burden among HIV-infected people in the United States using population-based data from HIV and cancer registries.
Methods
Overall Approach and Data Sources
In overview, we estimated national counts of cancers in HIV-infected individuals by applying cancer incidence rates for HIV-infected people to the US HIV population. We obtained excess counts by subtracting the expected number of cancers, calculated using general population cancer rates.
As described in the Supplementary Methods (available online), we derived cancer rates for the US HIV population from the HIV/AIDS Cancer Match (HACM) Study, a linkage of population-based state HIV and cancer registries (http://www.hivmatch.cancer.gov) (14). Only anonymized data were retained by investigators. The HACM Study was approved by institutional review boards and, as required, at participating registries.
We identified cancers in HIV-infected people from six states in the HACM Study using linked cancer registry data, and counted the first instance of each cancer type. We evaluated 11 cancer outcomes: the three ADCs, five NADCs that occur at elevated rates in HIV-infected people (Hodgkin’s lymphoma [HL] and cancers of the lung, anus, liver, and oral cavity/pharynx), and three cancers common in the general population (prostate, colorectum, and female breast). We grouped the remaining cancers into a “miscellaneous” category. KS, NHL, and cervical cancer occurring as AIDS-defining events were classified as occurring among people with HIV-only (ie, people who had not developed AIDS).
We corrected for two known errors in the HACM data (see Supplementary Methods, available online). First, we corrected for outmigration that prevents cancer ascertainment by decreasing person-time by 10% for HIV-infected people assessed 10 or more years since being documented to live in their registry area (15). Second, we increased cancer rates for HIV-infected people by 23.3% to adjust for imperfect sensitivity of the HIV cancer registry match (16).
We obtained 2010 US general population cancer rates from 18 Surveillance, Epidemiology, and End Results (SEER) cancer registries. For KS, we instead used 1973 to 1979 SEER rates, because subsequent rates were strongly influenced by HIV-infected cases (17).
We applied cancer rates to CDC surveillance estimates of the number of individuals living with diagnosed HIV at mid-year (ie, July 1) 2010 (1). Data reported through June 2012 to the CDC’s National HIV Surveillance System were used to generate estimates stratified by calendar year, age at year-end, transmission category (ie, sex and HIV risk group), race/ethnicity, and time since AIDS. These estimates adjusted for reporting delays and missing transmission category but not incomplete reporting. To obtain 2010 mid-year estimates, we averaged 2009 and 2010 end-of-year counts.
Statistical Analysis
To estimate 2010 cancer incidence rates for HIV-infected people, we fit Poisson regression models to 2003–2010 HACM data separately for each cancer type. Models included calendar year (as a continuous variable) and adjustment for age, sex, race/ethnicity, HIV risk group, time since AIDS, and registry area (see categories in Table 1 and additional details in the Supplementary Methods, available online). Using these models, we predicted 2010 cancer rates specific to age, sex, race/ethnicity, HIV risk group, time since AIDS, and registry area. We collapsed the six registry-specific incidence rates by weighting estimates based on registry size (see Supplementary Methods, available online).
Table 1.
Demographic and HIV-related characteristics of people living with diagnosed HIV infection based on CDC surveillance estimates (mid-year 2010) and person-time in the HACM study (2003–2010)*
| Characteristic | CDC, mid-year 2010 | HACM Study, 2003–2010 | ||
|---|---|---|---|---|
| No. people | Percentage | No. person-years | Percentage | |
| Total | 859 522 | 100 | 825 415 | 100 |
| Age, y | ||||
| 0–14 | 4307 | 0.5 | 7072 | 0.9 |
| 15–29 | 86 287 | 10.0 | 77 569 | 9.4 |
| 30–39 | 163 945 | 19.1 | 209 851 | 25.4 |
| 40–49 | 312 526 | 36.4 | 324 952 | 39.4 |
| 50–59 | 216 430 | 25.2 | 161 154 | 19.5 |
| 60–69 | 62 728 | 7.3 | 37 087 | 4.5 |
| ≥70 | 13 300 | 1.5 | 7730 | 0.9 |
| Sex | ||||
| Male | 643 627 | 74.9 | 613 835 | 74.4 |
| Female | 215 896 | 25.1 | 211 581 | 25.6 |
| HIV risk group | ||||
| MSM | 478 956 | 55.7 | 370 179 | 44.8 |
| Male heterosexual | 69 342 | 8.1 | 48 054 | 5.8 |
| Female heterosexual | 153 950 | 17.9 | 94 432 | 11.4 |
| Male IDU | 87 037 | 10.1 | 95 989 | 11.6 |
| Female IDU | 54 412 | 6.3 | 53 271 | 6.5 |
| Other/unknown† | 15 826 | 1.8 | 163 490 | 19.8 |
| Time since an AIDS diagnosis, mo | ||||
| HIV-only | 385 221 | 44.8 | 333 386 | 40.4 |
| 1–3 | 6371 | 0.7 | 9897 | 1.2 |
| 4–24 | 48 232 | 5.6 | 65 825 | 8.0 |
| 25–60 | 83 774 | 9.7 | 106 941 | 13.0 |
| >60 | 335 925 | 39.1 | 309 367 | 37.5 |
| Race/ethnicity | ||||
| Non-Hispanic white | 291 921 | 34.0 | 274 913 | 33.3 |
| Non-Hispanic black | 377 094 | 43.9 | 380 372 | 46.1 |
| Hispanic/Latino | 161 969 | 18.8 | 158 129 | 19.2 |
| Other/unknown | 28 540 | 3.3 | 12 001 | 1.5 |
* CDC mid-year 2010 counts were obtained by averaging 2009 and 2010 end-of-year counts. Sums within a category may not exactly equal the overall total because of rounding. Individuals who are both men who have sex with men (MSM) and injection drug users are classified as MSM. CDC = Centers for Disease Control and Prevention; HACM = HIV/AIDS Cancer Match; IDU = injection drug users; MSM = men who have sex with men.
† CDC imputes HIV risk groups for people with unknown risk group status, thereby assigning them to other groups.
To estimate total cancers among HIV-infected people, we applied these rates to the 2010 mid-year HIV population stratified by age, sex, race/ethnicity, HIV risk group, and time since AIDS. Separately, we estimated expected cancers by applying SEER rates, stratified by age, sex, and race/ethnicity, to the same population.
We estimated excess cancers by subtracting expected cancers from total cancers. For individual cancer types or demographic groups, we calculated the percentage of total cancers that were excess (“percentage excess”) as (total – expected total)/total. The percentage excess can be approximated as (SIR – 1)/SIR, where the SIR is the standardized incidence ratio, which is often used to quantify the relative risk of cancer in HIV-infected people. Where deficits occurred, we calculated the percentage of expected cases that were in deficit (“percentage deficit”) as (total – expected total)/expected total.
We performed a parametric bootstrap procedure to obtain 95% confidence intervals and P values for equality of and trends in proportions. Equality P values were based on the empirical distribution of squared differences in proportions generated under the null hypothesis of no difference, while trend P values were derived from linear regression models fit to log-transformed estimates of the proportions (see Supplementary Methods, available online). A P value of less than .05 was considered statistically significant, and all statistical tests were two-sided.
We summed across stratified estimates of cancer counts to present total and excess cancers by cancer type and demographic subgroup. We also cross-classifed excess cancers by demographic category and cancer type.
Results
An estimated 859 522 people were living with diagnosed HIV infection in the United States at mid-year 2010 (Table 1). The majority were age 40 to 59 years, and 74.9% were male, predominantly men who have sex with men (MSM). While 44.8% had no prior AIDS diagnosis (ie, HIV-only), another 39.1% had been living five or more years since AIDS. As shown in Table 1, the distribution of person-time in the HACM Study (2003–2010), from which cancer rates for HIV-infected people were derived, was similar, except that a larger proportion of individuals in the HACM Study had other/unknown HIV risk group. This difference arises because the CDC imputes HIV risk groups for people of unknown risk group (1), thereby assigning them to other groups.
An estimated 7760 (95% confidence interval [CI] = 7330 to 8320) total cancers occurred among HIV-infected people in 2010 (Table 2). Approximately one-third (34%) were ADCs, and two-thirds (66%) were NADCs. The most common cancers were NHL (n = 1650), KS (n = 910), lung cancer (n = 840), and anal cancer (n = 760).
Table 2.
Estimated total and excess cancer cases among people living with HIV in the United States in 2010, by cancer type
| Cancer type | Estimated cases* | Percentage excess or deficit† (95% CI) | Percentage of total excess‡ (95% CI) | ||
|---|---|---|---|---|---|
| Total No. (95% CI) | Expected No. | Excess or deficit No. (95% CI) | |||
| AIDS-defining cancers | 2640 (2400 to 2890) | 230 | 2410 (2170 to 2660) | 91 (90 to 92) | 54 (49 to 58) |
| Non-Hodgkin’s lymphoma§ | 1650 (1470 to 1840) | 200 | 1440 (1270 to 1630) | 88 (86 to 89) | 32 (28 to 36) |
| Kaposi’s sarcoma | 910 (780 to 1060) | 2 | 910 (780 to 1060) | 100 (100 to 100) | 20 (18 to 23) |
| Cervix | 80 (50 to 120) | 30 | 50 (20 to 90) | 66 (46 to 78) | 1 (0 to 2) |
| Non–AIDS defining cancers | 5130 (4780 to 5580) | 3620 | 1510 (1170 to 1960) | 29 (24 to 35) | 46 (42 to 51) |
| Lung | 840 (710 to 1000) | 400 | 440 (300 to 600) | 52 (43 to 60) | 10 (7 to 13) |
| Anus | 760 (630 to 930) | 20 | 740 (610 to 910) | 97 (97 to 98) | 17 (14 to 19) |
| Prostate | 570 (450 to 720) | 970 | -390 (-520 to -250) | -41 (-53 to -26) | 0 (0 to 0) |
| Liver | 390 (300 to 500) | 110 | 280 (200 to 390) | 73 (65 to 79) | 6 (4 to 9) |
| Colorectum | 360 (270 to 470) | 380 | -20 (-110 to 90) | -6 (-28 to 20) | 0 (0 to 2) |
| Hodgkin’s lymphoma | 320 (240 to 420) | 30 | 290 (210 to 390) | 91 (88 to 93) | 6 (5 to 9) |
| Oral cavity and pharynx | 280 (210 to 380) | 140 | 140 (70 to 240) | 51 (34 to 64) | 3 (2 to 5) |
| Female breast | 180 (120 to 260) | 300 | -130 (-190 to -40) | -42 (-61 to -14) | 0 (0 to 0) |
| Miscellaneous | 1430 (1270 to 1620) | 1270 | 160 (-9 to 350) | 11 (-1 to 22) | 4 (0 to 8) |
| All cancers (total) | 7760 (7330 to 8320) | 3850 | 3920 (3480 to 4470) || | 50 (48 to 54) | 100 |
* Cancer cases are rounded to the nearest ten and may not always sum exactly.
† The percentage excess is calculated as excess/total; the percentage deficit is calculated as deficit/expected (see Methods).
‡ This calculation is restricted to the cancers that were present in excess. Deficit cancers were therefore set equal to zero, giving a denominator of 4460 cancer cases for the point estimates. For the groups of AIDS-defining and non–AIDS defining cancers, percentages are calculated as the sum of the percentages for individual cancer sites.
§ Only certain subtypes of non-Hodgkin’s lymphoma (NHL) are AIDS defining, but we grouped all NHLs together, as most cases among HIV-infected people are the AIDS-defining subtypes (40).
|| The excess displayed for all cancers is the net excess. Disregarding cancers occurring in deficit, the total excess is 4460 cancers.
After subtracting expected background cancers, a statistically significant excess occurred for NHL, KS, HL, and cervical, lung, anal, liver, and oral cavity/pharyngeal cancers (Table 2). Among the 4460 excess cancers, roughly half were ADCs (54%) and half NADCs (46%). The largest percentage excess was for KS (essentially all cases were excess), followed by anal cancer (97%), HL (91%), and NHL (88%). Seventy-nine percent of excess cases (n = 3530) were NHL, KS, anal cancer, or lung cancer. Fewer than expected cases occurred for prostate cancer (deficit of 41% of expected cases) and breast cancer (42%). Accounting for deficit cancers, the net excess was approximately half of total cancers (50%, n = 3920, 95% CI = 3480 to 4470) (Table 2).
The magnitude of the cancer burden varied across demographic groups (Table 3). Total cancer burden was highest among people age 50 to 59 years (n = 2540), MSM (n = 4540), those living five or more years with AIDS (n = 3410), and non-Hispanic blacks (n = 3290). The largest excess burden was among people age 40 to 49 years (n = 1610), though the largest percentage excess was among people age 15 to 29 years (93%). There was no excess among people age 70 years and older, among whom a small deficit (6%) was observed.
Table 3.
Estimated total and excess cancer cases among people living with HIV in the United States in 2010, by demographic and HIV-related characteristics
| Characteristic | Estimated cases* | Percentage excess or deficit† (95% CI) | Percentage of total excess‡ (95% CI) | ||
|---|---|---|---|---|---|
| Total No. (95% CI) | Expected No. | Excess or deficit No. (95% CI) | |||
| Age, y | |||||
| 0–14 | 4 (2 to 8) | 1 | 4 (1 to 8) | 87 (70 to 93) | 0 (0 to 0) |
| 15–29 | 350 (300 to 400) | 30 | 320 (280 to 370) | 93 (92 to 94) | 8 (7 to 9) |
| 30–39 | 970 (890 to 1080) | 140 | 830 (750 to 930) | 85 (84 to 87) | 21 (19 to 23) |
| 40–49 | 2370 (2200 to 2560) | 760 | 1610 (1450 to 1810) | 68 (66 to 70) | 41 (38 to 44) |
| 50–59 | 2540 (2370 to 2740) | 1590 | 950 (780 to 1150) | 37 (33 to 42) | 24 (21 to 27) |
| 60–69 | 1240 (1120 to 1380) | 1020 | 220 (100 to 360) | 18 (9 to 26) | 6 (3 to 8) |
| ≥70 | 290 (250 to 350) | 310 | -20 (-60 to 40) | -6 (-21 to 11) | 0 (0 to 1) |
| Sex | |||||
| Male | 6240 (5880 to 6710) | 3020 | 3210 (2850 to 3690) | 52 (49 to 55) | 82 (78 to 86) |
| Female | 1530 (1370 to 1710) | 830 | 700 (540 to 890) | 46 (40 to 52) | 18 (14 to 22) |
| HIV risk group | |||||
| MSM§ | 4540 (4260 to 4920) | 2000 | 2540 (2260 to 2920) | 56 (53 to 59) | 65 (61 to 69) |
| Male heterosexual | 730 (640 to 830) | 410 | 320 (230 to 420) | 44 (36 to 51) | 8 (6 to 10) |
| Female heterosexual | 980 (860 to 1120) | 570 | 410 (290 to 550) | 42 (34 to 49) | 10 (8 to 13) |
| Male IDU | 920 (840 to 1030) | 590 | 320 (240 to 430) | 35 (29 to 42) | 8 (7 to 10) |
| Female IDU | 520 (450 to 590) | 240 | 270 (210 to 350) | 53 (46 to 59) | 7 (5 to 9) |
| Other/unknown | 80 (70 to 90) | 30 | 50 (40 to 60) | 59 (55 to 64) | 1 (1 to 1) |
| Time since an AIDS diagnosis | |||||
| HIV-only | 2810 (2620 to 3060) | 1430 | 1380 (1190 to 1630) | 49 (45 to 53) | 35 (32 to 38) |
| 1–3 months | 240 (210 to 280) | 20 | 220 (190 to 260) | 91 (90 to 92) | 6 (5 to 6) |
| 4–24 months | 540 (480 to 600) | 160 | 370 (320 to 440) | 70 (66 to 73) | 10 (8 to 11) |
| 25–60 months | 760 (690 to 840) | 320 | 440 (370 to 520) | 58 (53 to 62) | 11 (10 to 13) |
| >60 months | 3410 (3170 to 3720) | 1910 | 1500 (1260 to 1800) | 44 (40 to 48) | 38 (35 to 42) |
| Race/ethnicity | |||||
| Non-Hispanic white | 2870 (2670 to 3130) | 1490 | 1380 (1180 to 1640) | 48 (44 to 52) | 35 (32 to 38) |
| Non-Hispanic black | 3290 (3080 to 3560) | 1820 | 1470 (1260 to 1740) | 45 (41 to 49) | 38 (34 to 41) |
| Hispanic/Latino | 1270 (1170 to 1390) | 450 | 820 (720 to 940) | 64 (61 to 67) | 21 (19 to 23) |
| Other/unknown | 330 (270 to 410) | 90 | 240 (180 to 320) | 74 (67 to 79) | 6 (5 to 8) |
* Cancer cases are rounded to the nearest ten and may not always sum exactly. IDU = injection drug users; MSM = men who have sex with men.
† The percentage excess is calculated as excess/total; the percentage deficit is calculated as deficit/expected (see Methods).
‡ Deficit cancers are set equal to zero to calculate percent of total excess.
§ Individuals who are both MSM and IDU are classified as MSM.
MSM represent 55.7% of the total HIV population (Table 1), and 65% of all excess cancers occurred in this group (n = 2540) (Table 3). Only 18% of excess cancers occurred among women (n = 700). Individuals living five or more years since AIDS experienced the largest absolute excess (n = 1500), followed by those with HIV-only (n = 1380), reflecting large population sizes in those groups (Table 1). However, the percentage excess was largest one to three months after AIDS (91%), declining to 44% for five or more years post-AIDS. Though most excess cancers occurred among non-Hispanic blacks (n = 1470) and whites (n = 1380), a larger percentage of cancers was excess among Hispanics/Latinos (64%) and the small group with other/unknown race/ethnicity (74%).
Cross-classifications of excess cancers by demographic category and cancer type are displayed graphically in Figures 1 and 2 (see also Supplementary Tables 1–4, available online). Most excess cases of NHL, KS, anal cancer, and lung cancer were in people age 30 to 59 years (Figure 1A). On a percentage scale (Figure 1B), among people age 15 to 29 years, 77% of excess cancers were ADCs. This percentage declined with age (P trend < .001), reaching 22% for people age 70 years and older, as the percentage represented by anal, lung, and liver cancers increased (5% in people age 15 to 29 years vs 55% in people age 70 years and older).
Figure 1.
Estimated excess cancer cases among people living with human immunodeficiency virus (HIV) in the United States in 2010, by age and time since an AIDS diagnosis. Excess cancer cases are displayed by age (A and B) and time since an AIDS diagnosis (C and D) on an absolute scale (A and C) and on a percentage scale (B and D). Totals for individual groups do not match Table 3 exactly, as deficits (ie, negative excesses) for individual cancer sites within each group are set equal to zero for graphical presentation. Estimates of excess cancers, including deficits, simultaneously stratified by cancer type and demographic and HIV-related characteristics, are presented in Supplementary Tables 1–4 (available online). HIV = human immunodeficiency virus; HL = Hodgkin’s lymphoma; KS = Kaposi’s sarcoma; NHL = non-Hodgkin’s lymphoma.
Figure 2.
Estimated excess cancer cases among people living HIV in the United States in 2010, by HIV risk group and race/ethnicity. Excess cancer cases are displayed by HIV risk group (A and B) and race/ethnicity (C and D) on an absolute scale (A and C) and on a percentage scale (B and D). Totals for individual groups do not match Table 3 exactly, as deficits (ie, negative excesses) for individual cancer sites within each group are set equal to zero for graphical presentation. Estimates of excess cancers, including deficits, simultaneously stratified by cancer type and demographic and HIV-related characteristics are presented in Supplementary Tables 1–4 (available online). Individuals who are both men who have sex with men (MSM) and injection drug users are classified as MSM. het. = heterosexuals; HIV = human immunodeficiency virus; IDU = injection drug users; KS = Kaposi’s sarcoma; MSM = men who have sex with men; NH = non-Hispanic; NHL = non-Hodgkin’s lymphoma.
A similar trend in cancer types occurred across time living with AIDS. While most excess cancers among those with HIV-only were ADCs (73%), the excess burden increasingly comprised NADCs with advancing time since AIDS (P trend < .001) (Figure 1D). Accordingly, while approximately half of excess KS and NHL cases occurred in people with HIV-only as AIDS-defining events (46% and 50%, respectively), most excess anal cancers (71%) and oral cavity/pharyngeal cancers (69%) occurred among those living five or more years with AIDS (Figure 1C).
Excess cancer types also varied across HIV risk and racial/ethnic groups. The majority of excess cancers were in MSM (Figure 2A): 88% of all excess KS cases, 83% of excess anal cancers, and 62% of excess NHLs occurred in MSM. Though injection drug users (IDUs) represented only 16.5% of the HIV population (Table 1), 46% of excess liver cancers and 42% of excess lung cancers occurred in IDUs (Figure 2A). Accordingly, for male and female IDUs combined, compared with MSM and heterosexuals combined, higher percentages of the excess burden were liver cancer (16% vs 4%, P < .001) and lung cancer (22% vs 7%, P < .001) (Figure 2B). While non-Hispanic whites and blacks experienced similar total numbers of excess cancers, (Figure 2C), a higher percentage was anal cancer in whites compared with blacks (24% vs 13%, P < .001) (Figure 2D) and a lower percentage was lung cancer (7% vs 13%, P < .001).
Discussion
In the US HIV population in 2010, there were an estimated 7760 total cancers, of which roughly half (n = 3920) were net excess cases, ie, cases above what would be expected based on general population rates. Among the excess cancers, approximately half were ADCs and half NADCs, and excess cases generally occurred in groups with large representation in the HIV population such as MSM. The excess cancer burden reflects elevated risks of certain cancers among HIV-infected people, and many cases could theoretically be prevented through targeted efforts directed at known risk factors.
Most cancers showing a high excess in Table 2 are caused by viruses (18). HIV-infected individuals have a high prevalence of many of these viruses and are likely to lose immune control of infections (3,19). Almost all cases of KS (caused by human herpesvirus 8) and anal cancer (human papillomavirus [HPV]) occurred in excess (18). Most NHL and HL cases among HIV-infected people are caused by Epstein-Barr virus (EBV) (18), and about 90% of both cancers were excess cases. Excesses in the range of 51% to 73% were observed for cervical cancer (for which HPV is a necessary cause), oral cavity/pharyngeal cancers (for which a subset is HPV-related), and liver cancer (which is often caused by hepatitis B or C viruses [HBV, HCV]) (18). Finally, much of the 52% excess of lung cancer is because of frequent tobacco use (20), although immunosuppression and/or inflammation may also be important (20,21).
With advancing age and time since AIDS, the excess cancer burden shifted from ADCs to NADCs (Figure 1). This pattern likely has several explanations. First, many ADCs occurred as AIDS-defining events, either in people with HIV-only or in the subsequent three-month period (where excess cases may represent delayed diagnoses of AIDS-defining events). Second, rates of successful HIV treatment, related to engagement in care and compliance with medication regimens, increase with age and time living with AIDS (22,23). These treatment trends contribute to a decline in the excess burden of ADCs, because HAART use reduces ADC risk (24,25). Finally, the shift to NADCs reflects prolonged exposure to risk factors such as smoking, HPV, and HCV. For example, anal cancer develops over a prolonged period of HPV infection and HIV-associated immunosuppression (26). Consistent with this model, we observed that over 70% of excess anal cancers occurred five or more years after AIDS diagnosis.
The concentration of cancer risk behaviors in certain groups explains some of our findings. Eighty-three percent of excess anal cancers occurred among MSM, who are at high risk for anal HPV infection. In contrast, IDUs experienced a disproportionate excess burden of liver and lung cancers. IDUs acquire HBV/HCV because of needle sharing, and nearly all IDUs are current or former smokers (20,27). By racial/ethnic group, there were more excess anal cancers in whites and more excess lung cancers in blacks. This likely reflects that most HIV-infected whites are MSM, while more blacks are IDUs or acquired HIV via heterosexual intercourse (1).
Our results quantify how the cancer burden in the HIV population differs from the general population and provide a framework for considering cancer control initiatives tailored to HIV-infected people. While the incidence of ADCs has declined, we highlight that 14 years after the introduction of HAART, over half of excess cancers were still ADCs. This continuing excess illustrates that improvements in HIV treatment at the population level must remain a priority. Implementing measures to promote access and adherence to HAART, especially targeted to young people and people with HIV-only, could prevent many excess ADCs. Indeed, while our results reflect an era when treatment for individuals with HIV-only was based on disease stage (ie, CD4 cell count), guidelines now recommend that all HIV-infected individuals be offered HAART (28,29).
Lung and anal cancers together represented 27% of the total excess, underscoring that the development of strategies for prevention and early detection of these cancers among HIV-infected people is warranted. These strategies might be targeted to subgroups that experience substantial NADC risk and excess burden. For example, anal Pap screening has been advocated for MSM (30,31), and while the utility of this approach is debated (32) our results highlight that targeting anal cancer prevention in older people and those with AIDS should be a priority. Additionally, HPV vaccination for HIV-infected individuals may aid prevention of anal, cervical, and oral cavity/pharyngeal cancers (33). However, vaccination confers less benefit to individuals with prior sexual partners, likely lessening overall effectiveness for the HIV population. IDUs are an appropriate population to prioritize for HCV testing and treatment (34) and screening for liver cancer (35). Heavy smokers (both IDUs and non-IDUs) might benefit from smoking cessation interventions or lung cancer screening (36).
We observed a 40% deficit for breast and prostate cancers, reflecting decreased incidence in HIV-infected people (9) and no excess of colorectal cancer. For prostate cancer, the deficit may partly reflect decreased prostate-specific antigen screening among HIV-infected people (37), while the cause of decreased breast cancer risk is unclear (38). Despite a lack of excess, the total burden of these cancers is substantial (n = 1110 or 14% of all cancers in 2010) and will increase as the HIV population ages and grows in size (7). Screening based on general population guidelines for these cancers is likely appropriate for HIV-infected people.
Our estimates apply to the entire US HIV population, and they are based on robust data sources for cancer rates and enumeration of the HIV population. One limitation is that people who are infected but not diagnosed with HIV (approximately 16% of the total HIV-infected population [39]) are not included, because precise counts of these individuals and their cancer rates are unknown. Therefore, these are underestimates of the true cancer burden in HIV-infected people. Further, our results are based on cancer rates for HIV-infected people derived from six state registries that are demographically similar to the overall 2010 US HIV population. Our analysis requires that stratum-specific rates from these registries represent cancer risk in the corresponding strata for the entire US HIV population, and large violations of this assumption, though unlikely, would invalidate our results. Additionally, our analysis synthesizes multiple data sources and could be limited by inconsistencies across them, and there is uncertainty related to use of modeling to derive rates. Finally, while our estimates provide one framework for tailoring cancer control strategies for HIV-infected people, additional factors should be considered, such as the absolute and relative risks for cancer and the potential risks and benefits of specific interventions.
Our study is the first to provide a national snapshot of cancer burden in HIV-infected people. Over time, the number and composition of these cancers will change with cancer incidence rates and the size and demographic makeup of the HIV population (7,8). The variation in the spectrum of excess cancers across groups reveals opportunities for cancer control initiatives targeted to HIV-infected people.
Funding
This research was supported by the Intramural Research Program of the National Cancer Institute at the National Institutes of Health. The following cancer registries were supported by the Surveillance, Epidemiology, and End Results Program of the National Cancer Institute: Connecticut (HHSN261201000024C) and New Jersey (HHSN261201300021I, N01-PC-2013-00021). The following cancer registries were supported by the National Program of Cancer Registries of the Centers for Disease Control and Prevention: Colorado (U58 DP000848-04), Georgia (5U58DP003875-01), Michigan (5U58DP000812-03), New Jersey (5U58/DP003931-02), and Texas (5U58DP000824-04). The New Jersey Cancer Registry was also supported by the state of New Jersey.
Supplementary Material
This work was presented at the 2014 Conference on Retroviruses and Opportunistic Infections (CROI).
The views expressed in this paper are those of the authors and should not be interpreted to reflect the views or policies of the National Cancer Institute, the Centers for Disease Control and Prevention, HIV/AIDS or cancer registries, or their contractors. The sponsor reviewed and approved final submission but did not have a role in design and conduct of the study, the collection, management, analysis, or interpretation of the data, the preparation of the manuscript, nor the decision to submit for publication.
The authors gratefully acknowledge the support and assistance provided by individuals at the following state HIV/AIDS and cancer registries: Colorado, Connecticut, Georgia, Michigan, New Jersey, and Texas. We also thank Timothy McNeel at Information Management Services for programming support.
References
- 1. Centers for Disease Control and Prevention. HIV Surveillance Report Vol. 23. Diagnoses of HIV infection in the United States and Dependent Areas, 2011. February 28, 2013.
- 2. Grulich AE, van Leeuwen MT, Falster MO, Vajdic CM. Incidence of cancers in people with HIV/AIDS compared with immunosuppressed transplant recipients: a meta-analysis. Lancet. 2007;370(9581):59–67. [DOI] [PubMed] [Google Scholar]
- 3. Guiguet M, Boue F, Cadranel J, et al. Effect of immunodeficiency, HIV viral load, and antiretroviral therapy on the risk of individual malignancies (FHDH-ANRS CO4): a prospective cohort study. Lancet Oncol. 2009;10(12):1152–1159. [DOI] [PubMed] [Google Scholar]
- 4. Engels EA. Non-AIDS-defining malignancies in HIV-infected persons: etiologic puzzles, epidemiologic perils, prevention opportunities. AIDS. 2009;23(8):875–885. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Palella FJ, Jr., Delaney KM, Moorman AC, et al. Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. HIV Outpatient Study Investigators. N Engl J Med. 1998;338(13):853–860. [DOI] [PubMed] [Google Scholar]
- 6. Sterne JA, Hernan MA, Ledergerber B, et al. Long-term effectiveness of potent antiretroviral therapy in preventing AIDS and death: a prospective cohort study. Lancet. 2005;366(9483):378–384. [DOI] [PubMed] [Google Scholar]
- 7. Shiels MS, Pfeiffer RM, Gail MH, et al. Cancer burden in the HIV-infected population in the United States. J Natl Cancer Inst. 2011;103(9):753–762. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Robbins HA, Shiels MS, Pfeiffer RM, Engels EA. Epidemiologic contributions to recent cancer trends among HIV-infected people in the United States. AIDS. 2014;28(6):881–890. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Howlader N, Noone AM, Krapcho M, et al. SEER Cancer Statistics Review, 1975–2010, National Cancer Institute. Bethesda, MD: http://seer.cancer.gov/csr/1975_2010/. April 1, 2013. [Google Scholar]
- 10. de Martel C, Ferlay J, Franceschi S, et al. Global burden of cancers attributable to infections in 2008: a review and synthetic analysis. Lancet Oncol. 2012;13(6):607–615. [DOI] [PubMed] [Google Scholar]
- 11. Schutze M, Boeing H, Pischon T, et al. Alcohol attributable burden of incidence of cancer in eight European countries based on results from prospective cohort study. BMJ. 2011;342:d1584. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Renehan AG, Soerjomataram I, Tyson M, et al. Incident cancer burden attributable to excess body mass index in 30 European countries. Int J Cancer. 2010;126(3):692–702. [DOI] [PubMed] [Google Scholar]
- 13. Gopal S, Achenbach CJ, Yanik EL, et al. Moving forward in HIV-associated cancer. J Clin Oncol. 2014;32(9):876–880. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Engels EA, Biggar RJ, Hall HI, et al. Cancer risk in people infected with human immunodeficiency virus in the United States. Int J Cancer. 2008;123(1):187–194. [DOI] [PubMed] [Google Scholar]
- 15. Simard EP, Pfeiffer RM, Engels EA. Spectrum of cancer risk late after AIDS onset in the United States. Arch Intern Med. 2010;170(15):1337–1345. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Shiels MS, Pfeiffer RM, Hall HI, et al. Proportions of Kaposi sarcoma, selected non-Hodgkin lymphomas, and cervical cancer in the United States occurring in persons with AIDS, 1980–2007. JAMA. 2011;305(14):1450–1459. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Chaturvedi AK, Mbulaiteye SM, Engels EA. Underestimation of relative risks by standardized incidence ratios for AIDS-related cancers. Ann Epidemiol. 2008;18(3):230–234. [DOI] [PubMed] [Google Scholar]
- 18. IARC Working Group on the Evaluation of Carcinogenic Risks to Humans. A review of human carcinogens. Part B: Biological agents. 2009. Lyon, France. [Google Scholar]
- 19. Biggar RJ, Chaturvedi AK, Goedert JJ, Engels EA. AIDS-related cancer and severity of immunosuppression in persons with AIDS. J Natl Cancer Inst. 2007;99(12):962–972. [DOI] [PubMed] [Google Scholar]
- 20. Kirk GD, Merlo CA. HIV infection in the etiology of lung cancer: confounding, causality, and consequences. Proc Am Thorac Soc. 2011;8(3):326–332. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Engels EA. Inflammation in the development of lung cancer: epidemiological evidence. Expert Rev Anticancer Ther. 2008;8(4):605–615. [DOI] [PubMed] [Google Scholar]
- 22. Hall HI, Gray KM, Tang T, et al. Retention in care of adults and adolescents living with HIV in 13U.S. areas. J Acquir Immune Defic Syndr. 2012;60(1):77–82. [DOI] [PubMed] [Google Scholar]
- 23. Hall HI, Frazier EL, Rhodes P, et al. Differences in human immunodeficiency virus care and treatment among subpopulations in the United States. JAMA Intern Med. 2013;173(14):1337–1344. [DOI] [PubMed] [Google Scholar]
- 24. International Collaboration on HIV and Cancer. Highly active antiretroviral therapy and incidence of cancer in human immunodeficiency virus-infected adults. J Natl Cancer Inst. 2000;92(22):1823–1830. [DOI] [PubMed] [Google Scholar]
- 25. Chen YC, Li CY, Liu HY, et al. Effect of antiretroviral therapy on the incidence of cervical neoplasia among HIV-infected women: a population-based cohort study in Taiwan. AIDS. 2014;28(5):709–715. [DOI] [PubMed] [Google Scholar]
- 26. Bertisch B, Franceschi S, Lise M, et al. Risk factors for anal cancer in persons infected with HIV: a nested case-control study in the Swiss HIV Cohort Study. Am J Epidemiol. 2013;178(6):877–884. [DOI] [PubMed] [Google Scholar]
- 27. Marshall MM, Kirk GD, Caporaso NE, et al. Tobacco use and nicotine dependence among HIV-infected and uninfected injection drug users. Addict Behav. 2011;36(1–2):61–67. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Panel on Antiretroviral Guidelines for Adults and Adolescents. Guidelines for the use of antiretroviral agents in HIV-1-infected adults and adolescents. U.S. Department of Health and Human Services. February 13, 2013.
- 29. Gunthard HF, Aberg JA, Eron JJ, et al. Antiretroviral treatment of adult HIV infection: 2014 recommendations of the International Antiviral Society-USA Panel. JAMA. 2014;312(4):410–425. [DOI] [PubMed] [Google Scholar]
- 30. Goldie SJ, Kuntz KM, Weinstein MC, et al. The clinical effectiveness and cost-effectiveness of screening for anal squamous intraepithelial lesions in homosexual and bisexual HIV-positive men. JAMA. 1999;281(19):1822–1829. [DOI] [PubMed] [Google Scholar]
- 31. Chiao EY, Giordano TP, Palefsky JM, Tyring S, El SH. Screening HIV-infected individuals for anal cancer precursor lesions: a systematic review. Clin Infect Dis. 2006;43(2):223–233. [DOI] [PubMed] [Google Scholar]
- 32. Machalek DA, Poynten M, Jin F, et al. Anal human papillomavirus infection and associated neoplastic lesions in men who have sex with men: a systematic review and meta-analysis. Lancet Oncol. 2012;13(5):487–500. [DOI] [PubMed] [Google Scholar]
- 33. Panel on Opportunistic Infections in HIV-Infected Adults and Adolescents. Guidelines for the prevention and treatment of opportunistic infection in HIV-infected adults and adolescents: recommendations from the Centers for Disease Control and Prevention, the National Institutes of Health, and the HIV Medicine Association of the Infectious Diseases Society of America. July 23, 2013.
- 34. Joshi D, O’Grady J, Dieterich D, Gazzard B, Agarwal K. Increasing burden of liver disease in patients with HIV infection. Lancet. 2011;377(9772):1198–1209. [DOI] [PubMed] [Google Scholar]
- 35. Bruno R, Puoti M, Sacchi P, et al. Management of hepatocellular carcinoma in human immunodeficiency virus-infected patients. J Hepatol. 2006;44(1 Suppl):S146–S150. [DOI] [PubMed] [Google Scholar]
- 36. Aberle DR, Adams AM, Berg CD, et al. Reduced lung-cancer mortality with low-dose computed tomographic screening. N Engl J Med. 2011;365(5):395–409. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Shiels MS, Goedert JJ, Moore RD, Platz EA, Engels EA. Reduced risk of prostate cancer in U.S. men with AIDS. Cancer Epidemiol Biomarkers Prev. 2010;19(11):2910–2915. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Goedert JJ, Schairer C, McNeel TS, et al. Risk of breast, ovary, and uterine corpus cancers among 85,268 women with AIDS. Br J Cancer. 2006;95(5):642–648. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Centers for Disease Control and Prevention. Monitoring selected national HIV prevention and care objectives by using HIV surveillance data - United States and 6 dependent areas - 2011. HIV Surveillance Supplemental Report 2013. 18[5]. October 1, 2013.
- 40. Shiels MS, Engels EA, Linet M, et al. The epidemic of non-Hodgkin lymphoma in the United States: disentangling the effect of HIV, 1992–2009. Cancer Epidemiol Biomarkers Prev. 2013;22(6):1069–1078. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.